Two analog counters for neural network implementation

Simple analog circuits which are useful for the implementation of the synchronous Boltzmann machine learning algorithms are presented. A simple charge-transfer-based analog counter is described. The authors give a functional model of its behavior and analyze the differences between this model and the counter implementation. They also present simulation results and the test of a prototype. Along the same lines, they study a switched-current-based counter, which achieves better results (dynamic range, linearity) through higher complexity     

A positively self-feedbacked Hopfield neural network architecture for crossbar switching

We propose a positively self-feedbacked Hopfield neural network architecture for efficiently solving crossbar switch problem. A binary Hopfield neural network architecture with additional positive self-feedbacks and its collective computational properties are studied. It is proved theoretically and confirmed by simulating the randomly generated Hopfield neural network with positive self-feedbacks that the emergent collective properties of the original Hopfield neural network also are present in this network architecture. The network architecture is applied to crossbar switching and results of computer simulations are presented and used to illustrate the computation power of the network architecture. The simulation results show that the Hopfield neural network architecture with positive self-feedbacks is much better than the previous works including the original Hopfield neural network architecture, Troudet's architecture and maximum neural network for crossbar switching in terms of both the computation time and the solution quality.     

权重连接神经网络的光电实现

在Ｈｏｐｆｉｅｌｄ模型基础上，对具有权重连接的Ｈｏｐｆｉｅｌｄ模型引入连接权重矩阵，这样只要在Ｈｏｐｆｉｅｌｄ内容寻址记忆光电阵更前多加一个连接权重矩阵阵列，则得具有权重连接的神经网络模型的光电实现。     

ANALOG CIRCUIT IMPLEMENTATION OF NEURAL NETWORK WITH HIGH PRECISION WEIGHTS

A current-mode MOS neuron circuit with 4-bit programmable weights is presented by using CMOS technology. The weights of the neurcn have high resolution and also can easily be digitally stored. The resolution can be extended into high levels such as 8-bit, etc. by the design methodology presented in this paper. The operational principle of the neuron is discussed. Circuit simulation has been made by use of SPICE II. The results give a good agreement for the design requirements.     

Intelligent switching control of a pneumatic muscle robot arm using learning vector quantization neural network

Pneumatic cylinders are one of the low-cost actuation sources used in industrial and prosthetic application, since they have a high power/weight ratio, high-tension force and long durability. However, problems with the control, oscillatory motion and compliance of pneumatic systems have prevented their widespread use in advanced robotics. To overcome these shortcomings, a number of newer pneumatic actuators have been developed, such as the McKibben Muscle, Rubber Actuator and Pneumatic Artificial Muscle (PAM) Manipulators. In this paper, the solution for position control of a robot arm with slow motion driven by two pneumatic artificial muscles is presented. However, some limitations still exist, such as a deterioration of the performance of transient response due to the changes in the external load. To overcome this problem, a switching algorithm of the control parameter using a learning vector quantization neural network (LVQNN) is proposed in this paper. The LVQNN estimates the external load of the pneumatic artificial muscle manipulator. The effectiveness of the proposed control algorithm is demonstrated through experiments with different external working loads.     

Analogue implementation of a neural network controller for UPSinverter applications

An analogue neural-network controller for UPS inverter applications is presented. The proposed neural-network controller is trained off-line using patterns obtained from a simulated controller, which had an idealized load-current-reference. Simulation results show that the proposed neural-network controller can achieve low total harmonic distortion under nonlinear loading condition and good dynamic responses under transient loading condition. To verify the performance of the proposed NN controller, a hardware inverter with an analogue neural network (NN) controller (using mainly operational amplifiers and resistors) is built. Additionally, for comparison purposes, a PI controller with optimized parameters is built. Experimental results confirm the simulation results and show the superior performance of the NN controller especially under rectifier-type loading condition. Implementing the analogue neural-network controller using programmable integrated circuits is also discussed     

Acknowledgment schemes for window flow control in a quasi-M/D/1cut-through switching network

Window flow control in a quasi-M/D/1 cut-through switching network with noisy channel is discussed. Two acknowledgment schemes for window edge control at the entry node are proposed, and their various properties investigated     

基于神经网络的功率模块开关损耗预测

功率半导体器件的开关损耗的准确预测和评估对研究系统设计、选择合适的散热系统和提高系统的可靠性都是很重要的。采用功率半导体器件的开关动态测试系统,可以自动调整直流母线电压、集电极电流、门极驱动电压和开关频率,记录开关动态电压、电流波形,计算获得大量的损耗数据。通过建立人工神经网络模型对开关损耗预测,并改变模型参数进行分析与研究,获取最佳模型。     

IP交换中流分类的神经网络方法

本文对IP交换机中流的分类问题进行了探讨。首先对现有的X／Y分类器进行了分析,指出其不足之处,然后总结了进行流判断时应考虑的各种因素。之后提出了一种利用Hopfield神经网络进行流分类的方法,并对其参数的选取进行了讨论。仿真结果表明,神经网络分类器可根据网络中流的具体情况,自动调节分类阈值,保证IP交换机利用的VC数基本稳定。与常规X／Y分类器相比,神经网络分类器可利用更少的VC交换更多的数据包。     

Temperature control with a neural fuzzy inference network

Although multilayered backpropagation neural networks (BPNNs) have demonstrated high potential in adaptive control, their long training time usually discourages their applications in industry. Moreover, when they are trained online to adapt to plant variations, the over-tuned phenomenon usually occurs. To overcome the weakness of the BPNN, we propose a neural fuzzy inference network (NFIN) suitable for adaptive control of practical plant systems in general and for adaptive temperature control of a water bath system in particular. The NFIN is inherently a modified Takagi-Sugeno-Kang (TSK)-type fuzzy rule based model possessing a neural network's learning ability. In contrast to the general adaptive neural fuzzy networks, where the rules should be decided in advance before parameter learning is performed, there are no rules initially in the NFIN. The rules in the NFIN are created and adapted as online learning proceeds via simultaneous structure and parameter identification. The NFIN has been applied to a practical water bath temperature control system. As compared to the BPNN under the same training procedure, the simulated results show that not only can the NFIN greatly reduce the training time and avoid the over-tuned phenomenon, but the NFIN also has perfect regulation ability. The performance of the NFIN is compared to that of the traditional PID controller and fuzzy logic controller (FLC) on the water bath temperature control system. The three control schemes are compared, with respect to set point regulation, ramp-point tracking, and the influence of unknown impulse noise and large parameter variation in the temperature control system. The proposed NFIN scheme has the best control performance     

Adaptive fuzzy neural network control for motor-toggle servomechanism

In this study, the dynamic responses of an adaptive fuzzy neural network (FNN) controlled toggle mechanism is described. The toggle mechanism is driven by a permanent magnet (PM) synchronous servo motor. First, based on the principle of computed torque, an adaptive controller is developed to control the position of a slider of the motor-toggle servomechanism. Since the selection of control gain of the adaptive controller has a significant effect on the system performance, an adaptive FNN controller is proposed to control the motor-toggle servomechanism. In the proposed adaptive FNN controller, an FNN is adopted to facilitate the adjustment of control gain on line. Moreover, simulated and experimental results due to a periodic sinusoidal command show that the dynamic behaviors of the proposed adaptive and adaptive FNN controllers are robust with regard to uncertainties.     

二级倒立摆的自适应神经网络控制

倒立摆系统是一种典型的非线性、多变量、不稳定系统,目前,对于这种复杂对象的控制问题在控制领域具有十分重要的研究价值。针对此种非线性系统的控制问题,提出一种智能控制方法来解决这个问题。通过应用神经网络控制和模糊控制相结合的方式,集合二者的优点,提出一种将BP算法与最小二乘算法相结合的算法,对Takagi-Sugeno模糊推理系统中的参数进行优化修正,设计一种自适应神经网络的模糊推理系统来控制倒立摆,实验结果证明该理论是准确可行的,与LQR实时控制相比响应速度快、精度高。     

An all-optical implementation of a 3-D crossover switching network

One of the more promising interconnection schemes proposed for use in photonic switching networks is the crossover interconnection network; however, reported implementations of the crossover have been limited in size and complexity. A large-scale cascadable implementation of the optical crossover network that capitalizes on planar symmetric self electrooptic effect device (S-SEED) arrays is discussed. A fully functional experimental prototype with 32 inputs and 32 outputs that was operated at a maximum rate of 55.7 kb/s is also discussed. It is also shown that S-SEED arrays can be operated as simple two-input two-output nodes (called 2-modules) within a controllable network     

Holographic implementation of a fully connected neural network

A holographic implementation of a fully connected neural network is presented. This model has a simple structure and is relatively easy to implement, and its operating principles and characteristics can be extended to other types of networks, since any architecture can be considered as a fully connected network with some of its connections missing. The basic principles of the fully connected network are reviewed. The optical implementation of the network is presented. Experimental results which demonstrate its ability to recognize stored images are given, and its performance and analysis are discussed based on a proposed model for the system. Special attention is focused on the dynamics of the feedback loop and the tradeoff between distortion tolerance and image-recognition capability of the associative memory     

Speed sensorless control of a bearingless induction motor with combined neural network and fractional sliding mode

To address the problem of speed and flux observation in sensorless control of a bearingless induction motor under the influence of parameter changes and external disturbances, a speed sensorless control strategy combining radial basis function (radial basis function, RBF) neural network and fractional sliding mode is proposed. According to the current error, fractional sliding mode control rate is designed to reduce the speed-observed chatter of the bearingless induction motor and its adverse effect on the rotor suspension stability. Then, combined with the theory of RBF neural network, the new optimal control rate is obtained by using its approximation ability. At the same time, the stability of two control rate is proved. Thus, the flux linkage and speed under normal operation, parameter change and external disturbance are observed and the new speed sensorless control is realized. The simulation and experimental results show that the proposed joint RBF neural network approximation algorithm and fractional sliding mode speed sensorless control system of the bearingless induction motor can not only effectively identify the flux and speed under three conditions of no-load, load disturbance and speed change, but also ensure the good suspension of the motor rotor in the x-axis and y-axis directions.     

FPGA implementation of high-speed neural network for power amplifier behavioral modeling

In this paper, a high-speed pipelined architecture of dynamic neural network is proposed for power amplifier behavioral modeling. This architecture is implemented on field programmable gate array (FPGA) using Xilinx system generator and Virtex-6 FPGA ML605 Evaluation Kit. The novelty of the proposed architecture is that it provides higher operating frequency, lower output latency, and less required resources. These improvements are obtained by reducing the bit-width data and by efficiently redistributing the inserted pipelining delays. The new pipelined architecture is evaluated and compared to the conventional and pseudo-conventional ones in terms of the resource utilization, the maximum operating frequency, and the modeling performances using the 16-QAM modeled test signal. This architecture is verified using JTAG hardware co-simulation both for single step and free-running clock modes.     

DSP-based multi-purpose control system for laser processing

In modern laser processing[1],the operations per-formed by a motion control system have to be done inreal ti me at high frequency. While PC hardware andsoftware are not opti mizedfor these kinds of operations ,the DSPandthe software onthe motioncontrol sy…     

Multistage network with globally controlled switching stages and its implementation using optical multi-interconnection modules

Plane-to-plane guided-wave-based interconnection modules are proposed as building blocks for scalable optoelectronic multistage interconnection networks (MINs). This approach leads naturally to a MIN paradigm based not on cascading switching stages containing several size-reduced crossbars, as in the shuffle-exchange (SE) networks, but on cascading permutation-reduced crossbars instead, one per stage. The interest of such an architecture lies in the control simplicity and scalability potential. Transparent circuit switching for permutation routing is possible in such an unbuffered "globally switched" multistage interconnection network (GSMIN). Preliminary experiments using fiber-based interconnection modules are presented. Performance analysis and simulation of a buffered GSMIN is also studied for packet routing purposes.     

神经网络下汽车控制系统故障诊断

汽车的自动化水平是不断提高的,传统维修模式依靠专业人员的经验和能力,这对于过于复杂的系统效果不佳。同时为了提高诊断的效率和准确性,也需要采用智能化、自动化的方案。人工神经网络有良好的自学习,自适应特性,适用于非线性映射,能够克服获取知识的困难,以及知识存储跟运行速度的矛盾问题,因此可以用来开发汽车复杂控制系统的故障诊断问题。诊断过程首先归一化故障数据,设计良好的神经网络收敛算法,对关键的汽车控制系统故障模式学习,将训练好的网络用于汽车故障诊断,弥补人工经验方法的不足。     

A DSP-based implementation of a new nonlinear control for a three-phase neutral point clamped boost rectifier prototype

This work presents the design and implementation of a nonlinear control strategy for a three-phase three-level neutral-point-clamped boost rectifier. The adopted control consists of nonlinear feedback linearization technique. The nonlinear state-space model of the rectifier was obtained in the dq0 reference frame. The input/output feedback linearization is then applied and the linearizing control law is derived. Therefore, the resulting model is linearized and decoupled in three independent subsystems. Afterwards, the stabilizing controllers are designed based on linear techniques to control line currents, output, and neutral point voltages. The control law is designed using Simulink/Matlab and applied to the converter via a 1.8-kHz pulsewidth modulator (PWM). Both control law and PWM signals are executed in real time using the DS1104 DSP of dSPACE. A 1.2 kW laboratory prototype is built for validation purposes. The proposed control law robustness is validated for diverse severe load and system parameter variations. It shows robust performance in terms of high power factor, low total harmonic distortion and output voltage ripples, small overshoot, and short settling time.